Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed FBox industrial IoT MCP helper with expected API-key and network use, but it should only be used by authorized operators of the connected equipment.
Install this only for an FBox account you are authorized to operate. Treat the API key, device network settings, precise locations, PLC data, alarm data, and VNC sessions as sensitive operational information. Before any PLC write, alarm acknowledgment, or remote session, verify the exact device, site, current value/state, and intended maintenance context; also be aware that timestamps are displayed in Beijing time unless you adjust expectations externally.
This markdown file instructs users to set and transmit an API key in environment variables and Authorization headers, but it does not warn users to keep the key secret, avoid sharing terminal history or screenshots, or rotate it if exposed. For a skill-installation guide that handles credentials, the omission is a user-safety disclosure gap under the markdown warning criterion.
The README advertises PLC point writes, alarm acknowledgments, and opening remote VNC sessions, but provides no safety caveats, authorization expectations, or operational guardrails. In an industrial IoT context, normalizing these actions as routine chat requests can lead users or agents to perform state-changing operations on live equipment without understanding consequences, increasing risk of process disruption or unsafe device behavior.
The trigger description includes broad terms such as device status, sensor data, alarms, remote monitoring, and VNC-related wording that may appear in ordinary industrial support conversations, causing the skill to activate when the user did not intend to operate industrial equipment tooling. In this context, unintended invocation is more dangerous than usual because the skill can access live OT/IoT data and supports sensitive actions such as alarm confirmation, PLC point writes, and remote VNC access, increasing the chance of confusing tool use or accidental operational impact.
Forcing all displayed times into Beijing time without user choice can mislead operators who are in other locales or who expect UTC/local plant time, especially when reviewing alarms, trends, or operational events. In an industrial operations setting, time misinterpretation can cause errors in incident correlation, shift handoff, and audit review, though this is more of a safety/usability risk than a direct security compromise.
This markdown file presents all operational instructions and parameter descriptions exclusively in Chinese. Under the language/locale policy rule, forcing a single language without user opt-in or documented justification is a natural-language policy concern.
The documented settings retrieval includes sensitive infrastructure details such as IP address, subnet mask, gateway, DNS, firmware version, network type, and dual-NIC information, but provides no warning or handling guidance for exposing this data. In an OT/industrial environment, these details materially aid network reconnaissance, targeting, and follow-on compromise of devices or connected systems.
The skill exposes precise device geolocation and detailed address fields, including both automatically reported and user-supplied coordinates, without any documented privacy notice, consent check, or data-minimization guidance. In an industrial IoT context, location data can reveal facility locations and operational footprints, increasing physical security and privacy risk if surfaced too broadly.
This markdown file contains user-facing natural-language instructions exclusively in Chinese, and nowhere indicates that the skill is region-specific or that users may choose another language. That can violate a language/locale policy requiring user opt-in or explicit justification for a forced language.
Lines L155-L156 state that after installing the skill, the assistant will automatically convert times to Beijing time (UTC+8). This is a locale behavior forced by the skill description, and the document does not offer user opt-in or explain that the skill is limited to a China-specific operational context.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
curl https://<fboxmcp-host>/health
返回 200 OK 表示服务正常。
All user-facing instructions in the file are in Chinese, and the document does not indicate that this locale is optional, region-specific, or that alternative language support exists. This can violate a language/locale policy when a skill forces a specific language without user opt-in.
The README is entirely in Chinese and all invocation examples are presented only in Chinese, with no indication that users may interact in other languages or choose their preferred locale. This can constitute a language/locale policy issue when a skill appears to prescribe a single language without opt-in or justification.